More explicit casts to fix clang warnings (incomplete)

Author: sw

ggml.c

@@ -479,8 +479,8 @@ static void quantize_row_q4_0_reference(const float * restrict x, void * restric
             const float v0 = x[i*QK + l + 0]*id;
             const float v1 = x[i*QK + l + 1]*id;
 
-            const uint8_t vi0 = ((int8_t) (round(v0))) + 8;
-            const uint8_t vi1 = ((int8_t) (round(v1))) + 8;
+            const uint8_t vi0 = (int8_t)roundf(v0) + 8;
+            const uint8_t vi1 = (int8_t)roundf(v1) + 8;
 
             assert(vi0 >= 0 && vi0 < 16);
             assert(vi1 >= 0 && vi1 < 16);
@@ -747,8 +747,8 @@ void quantize_row_q4_1(const float * restrict x, void * restrict y, int k) {
             const float v0 = (x[i*QK + l + 0] - min)*id;
             const float v1 = (x[i*QK + l + 1] - min)*id;
 
-            const uint8_t vi0 = round(v0);
-            const uint8_t vi1 = round(v1);
+            const uint8_t vi0 = roundf(v0);
+            const uint8_t vi1 = roundf(v1);
 
             assert(vi0 >= 0 && vi0 < 16);
             assert(vi1 >= 0 && vi1 < 16);
@@ -2173,16 +2173,16 @@ inline static void ggml_vec_scale_f32(const int n, float * y, const float   v) {
 #endif
 }
 
-inline static void ggml_vec_norm_f32 (const int n, float * s, const float * x) { ggml_vec_dot_f32(n, s, x, x); *s = sqrt(*s);   }
+inline static void ggml_vec_norm_f32 (const int n, float * s, const float * x) { ggml_vec_dot_f32(n, s, x, x); *s = sqrtf(*s);   }
 inline static void ggml_vec_sqr_f32  (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = x[i]*x[i];   }
-inline static void ggml_vec_sqrt_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = sqrt(x[i]); }
+inline static void ggml_vec_sqrt_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = sqrtf(x[i]); }
 inline static void ggml_vec_abs_f32  (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = fabsf(x[i]); }
 inline static void ggml_vec_sgn_f32  (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = (x[i] > 0.f) ? 1.f : ((x[i] < 0.f) ? -1.f : 0.f); }
 inline static void ggml_vec_step_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = (x[i] > 0.f) ? 1.f : 0.f; }
 inline static void ggml_vec_relu_f32 (const int n, float * y, const float * x) { for (int i = 0; i < n; ++i) y[i] = (x[i] > 0.f) ? x[i] : 0.f; }
 
-static const float GELU_COEF_A    = 0.044715;
-static const float SQRT_2_OVER_PI = 0.79788456080286535587989211986876;
+static const float GELU_COEF_A    = 0.044715f;
+static const float SQRT_2_OVER_PI = 0.79788456080286535587989211986876f;
 
 inline static float ggml_gelu_f32(float x) {
     return 0.5f*x*(1.0f + tanhf(SQRT_2_OVER_PI*x*(1.0f + GELU_COEF_A*x*x)));
@@ -7565,8 +7565,8 @@ static void ggml_compute_forward_rope_f32(
                     const float * const src = (float *)((char *) src0->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
                           float * dst_data  = (float *)((char *)  dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
 
-                    double x0 = src[0];
-                    double x1 = src[1];
+                    double x0 = (double)src[0];
+                    double x1 = (double)src[1];
 
                     dst_data[0] = x0*cos_theta - x1*sin_theta;
                     dst_data[1] = x0*sin_theta + x1*cos_theta;
@@ -7621,8 +7621,8 @@ static void ggml_compute_forward_rope_f16(
                     const ggml_fp16_t * const src = (ggml_fp16_t *)((char *) src0->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
                           ggml_fp16_t * dst_data  = (ggml_fp16_t *)((char *)  dst->data + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
 
-                    double x0 = ggml_fp16_to_fp32(src[0]);
-                    double x1 = ggml_fp16_to_fp32(src[1]);
+                    double x0 = (double)ggml_fp16_to_fp32(src[0]);
+                    double x1 = (double)ggml_fp16_to_fp32(src[1]);
 
                     dst_data[0] = ggml_fp32_to_fp16(x0*cos_theta - x1*sin_theta);
                     dst_data[1] = ggml_fp32_to_fp16(x0*sin_theta + x1*cos_theta);
@@ -8298,7 +8298,7 @@ static void ggml_compute_forward_flash_attn_f32(
     const int ir0 = dr*ith;
     const int ir1 = MIN(ir0 + dr, nr);
 
-    const float scale = 1.0/sqrt((double) D);
+    const float scale = 1.0f/sqrtf(D);
 
     //printf("P=%d N=%d D=%d ir0=%d ir1=%d scale = %f\n", P, N, D, ir0, ir1, scale);
 
@@ -8507,7 +8507,7 @@ static void ggml_compute_forward_flash_attn_f16(
     const int ir0 = dr*ith;
     const int ir1 = MIN(ir0 + dr, nr);
 
-    const float scale = 1.0/sqrt((double) D);
+    const float scale = 1.0f/sqrtf(D);
 
     //printf("P=%d N=%d D=%d ir0=%d ir1=%d scale = %f\n", P, N, D, ir0, ir1, scale);
 

main.cpp

@@ -73,7 +73,7 @@ std::vector<double> softmax(const std::vector<float>& logits) {
     for (size_t i = 0; i < logits.size(); i++) {
         // Subtract the maximum logit value from the current logit value for numerical stability
         float logit = logits[i] - max_logit;
-        double exp_logit = std::exp(logit);
+        double exp_logit = std::exp((double)logit);
         sum_exp += exp_logit;
         probs[i] = exp_logit;
     }
@@ -375,10 +375,10 @@ int main(int argc, char ** argv) {
 
         if ((int) embd_inp.size() <= input_consumed && !is_interacting) {
             // out of user input, sample next token
-            const float top_k          = params.top_k;
-            const float top_p          = params.top_p;
-            const float temp           = params.temp;
-            const float repeat_penalty = params.repeat_penalty;
+            const int top_k             = params.top_k;
+            const double top_p          = (double)params.top_p;
+            const double temp           = (double)params.temp;
+            const double repeat_penalty = (double)params.repeat_penalty;
 
             llama_token id = 0;